9856–9873 Nucleic Acids Research, 2015, Vol. 43, No. 20 Published online 7 October 2015 doi: 10.1093/nar/gkv1026 Control of the localization and function of a miRNA silencing component TNRC6A by Argonaute protein Kenji Nishi1, Tomoko Takahashi1, Masataka Suzawa1, Takuya Miyakawa2, Tatsuya Nagasawa1, Yvelt Ming3, Masaru Tanokura2 and Kumiko Ui-Tei1,3,*
1Department of Biological Sciences, Graduate School of Science, University of Tokyo, Tokyo 113-0033, Japan, 2Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, University of Tokyo, Tokyo 113-8657, Japan and 3Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, University of Tokyo, Chiba-ken 277-8651, Japan
Received April 14, 2015; Revised September 25, 2015; Accepted September 28, 2015
ABSTRACT complexes, leading to translational repression and mRNA degradation of miRNA-targeted mRNAs (4,11–17). GW182 family proteins play important roles in mi- In human HEp-2, HeLa, and many other cell lines, a croRNA (miRNA)-mediated RNA silencing. They di- human GW182 family protein, TNRC6A, is known to rectly interact with Argonaute (Ago) proteins in pro- be localized mainly in the processing (P) bodies (18–20), cessing bodies (P bodies), cytoplasmic foci involved which are cytoplasmic foci that contain proteins involved in in mRNA degradation and storage. Recently, we re- mRNA degradation, storage, and translational repression vealed that a human GW182 family protein, TNRC6A, (reviewed in 21). However, recently we found a nuclear lo- is a nuclear-cytoplasmic shuttling protein, and its calization signal (NLS) and a nuclear export signal (NES) subcellular localization is regulated by its own nu- in the central region of TNRC6A and showed that it is a clear localization signal and nuclear export sig- nuclear–cytoplasmic shuttling protein and its subcellular nal. Regarding the further controlling mechanism localization is regulated by its own NLS and NES (10). In good agreement with our previous result, several of recent of TNRC6A subcellular localization, we found that reports have suggested that human GW182 proteins might TNRC6A protein is tethered in P bodies by direct in- function in the nucleus as well as in the cytoplasm. Chro- teraction with Ago2 under Ago2 overexpression con- matin silencing and alternative splicing, or transcriptional dition in HeLa cells. Furthermore, it was revealed that control targeting promoter RNA of inflammatory pathway such Ago proteins might be strongly tethered in the genes is shown to be associated with RNA silencing fac- P bodies through Ago-bound small RNAs. Thus, our tors in the nucleus (10,22–24). Furthermore, immunohisto- results indicate that TNRC6A subcellular localization chemical analyses of several human cancers, such as gas- is substantially controlled by the interaction with Ago tric, colorectal, prostate and esophageal cancers, show the proteins. Furthermore, it was also revealed that the strong nuclear localization of TNRC6A (25,26). Thus, the TNRC6A subcellular localization affects the RNA si- actual subcellular localization is different due to the cell lencing activity. types, and the molecular machinery involved in specifica- tion of the subcellular localization of TNRC6A remained unknown. INTRODUCTION In our previous report, we showed that TNRC6A pro- tein with NES mutations (TNRC6A-NES-mut) is predom- In small RNA-mediated gene silencing, Argonaute (Ago) inantly localized in the nucleus, due to the defect of the cy- proteins play key roles by directly binding to microRNAs toplasmic export machinery via its NES (10). Regarding (miRNAs) or small interfering RNAs (siRNAs) (reviewed the further controlling mechanism of TNRC6A subcellu- in 1,2). GW182 family proteins are shown to be necessary lar localization, in this paper, we show that, when wild-type for miRNA-mediated RNA silencing, hereafter referred to Ago proteins are overexpressed, TNRC6A-NES-mut pro- as miRNA silencing, in animals (reviewed in 3,4). They in- teins are tethered in P bodies by direct binding to abundant teract with Ago proteins via their N-terminal regions con- amount of Ago proteins, but they are not tethered by Ago2 taining multiple glycine-tryptophan (GW) repeats (5–10). mutant protein deficient for binding to small RNA and also On the other hand, their C-terminal regions, called as silenc- TNRC6A protein. These results suggest that TNRC6A pro- ing domain, recruit cytoplasmic poly(A)-binding protein 1 tein is tethered by the excessive amount of Ago proteins in (PABPC1), the PAN2-PAN3 and CCR4-NOT deadenylase
*To whom correspondence should be addressed. Tel: +81 3 5841 3043; Fax: +81 3 5841 3044; Email: [email protected]
C The Author(s) 2015. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] Nucleic Acids Research, 2015, Vol. 43, No. 20 9857 the P bodies through Ago-bound small RNAs, which is con- from pcDNA3.1-FLAG/HA-Dicer using primers (5 - sidered to form base-pairing with complementary RNAs in AAAAGGAAAAGCGGCCGCTCTAGAGGGCCCGT the P bodies. Our results strongly suggest that the nuclear- TTAAACCCG-3 and 5 - AAAGGGCTCGAGAGCG cytoplasmic shuttling of TNRC6A is mechanistically reg- TAATCGGGCACGTCATAAGG-3 ) for deleting Dicer ulated by its own NLS and NES, but its subcellular local- region. Both fragments were digested with XhoI and NotI, ization is substantially determined by the comparative ex- and ligated. For construction of FLAG/HA-RCK/p54 pression level of cellular Ago proteins. Furthermore, it was expression plasmid, pFLAG/HA-RCK/p54, a cDNA frag- revealed that the cytoplasmic TNRC6A protein enhances ment of RCK/p54 CDS region was amplified using cDNA miRNA silencing activity when excessive amount of Ago purified from HEK293 cells by primers (5 - AAAGGGAA proteins were expressed, although TNRC6A overexpression GCTTATGAGCACGGCCAGAACAGAGAAC-3 and repressed both RNA interference (RNAi) and miRNA si- 5 -AAAGGGTCTAGATTAAGGTTTCTCATCTTCTA lencing activity with low level of Ago2 protein in the cyto- CAGGC-3 ). Amplified RCK/p54 fragment was digested plasm. with HindIII and XbaI, and ligated with HindIII/XbaI- digested fragment of pcDNA3.1-FLAG/HA-Dicer. The expression constructs of chemokine (C-X-C mo- MATERIALS AND METHODS tif) receptor 4 (CXCR4) shRNA and control DsRed shRNA, named pSUPER-CXCR4 and pSUPER-DsRed, Plasmid construction respectively, were constructed as described previously The expression plasmids of pmyc-GFP, a full-length (10). The plasmid expressing Renilla luciferase har- TNRC6A, pmyc-GFP-TNRC6A, and its derivatives were boring a perfectly complementary target site or four constructed as described previously (10). pIRESneo- target sites with a central bulge for CXCR4 siRNA in FLAG/HA-Ago1, -Ago2, -Ago3, and -Ago4 (27)were its 3 UTR, psiCHECK-CXCR4-perfect-match (PM) or kindly provided by Dr Thomas Tuschl through Ad- psiCHECK-CXCR4-mismatch (MM) 4×, was constructed dgene, and designated as pFLAG/HA-Ago1, -Ago2, by inserting a pair of annealed oligonucleotides (5 - and -Ago3, respectively, in this study. For construction TCGAGAAGTTTTCACTCCAGCTAACAG-3 and 5 - of pFLAG/HA-FL, FL cDNA was amplified from AATTCTGTTAGCTGGAGTGAAAACTTC-3 )ortwo firefly luciferase gene in pGL3-Control (Promega) by pairs of annealed oligonucleotides to form a tandem inser- PCR using primers (5 - AAAAGGAAAAGCGGCC tion with four target sites (5 -TCGAGAAGTTTTCACA GCATGGAAGACGCCAAAAACATAAAG-3 and AAGCTAACAAGTCAAGTTTTCACAAAGCTAACA- 5 -AAAGGGGAATTCTTACACGGCGATCTTTCC 3 and 5 -TTGACTTGTTAGCTTTGTGAAAAC GCCCT-3 ). The amplified product was digested with TTGACTTGTTAGCTTTGTGAAAACTTC-3 , NotI and EcoRI, and ligated with NotI/EcoRI-digested and 5 -AGTCAAGTTTTCACAAAGCTAACAAG fragment of pFLAG/HA-Ago2. For construction of an TCAAGTTTTCACAAAGCTAACAG-3 and 5 - empty vector, pFLAG/HA, an untagged Ago2 expression AATTCTGTTAGCTTTGTGAAAACTTGACTTGT construct, pAgo2, and pFLAG/HA-Ago2-Y529E encod- TAGCTTTGTGAAAAC-3 ), respectively, into the ing Ago2 with a substitution from tyrosine to glutamic XhoI/EcoRI sites of psiCHECK-1 (Promega). acid at amino acid residue 529, the linear fragments with such deletions or mutations were amplified from Cell culture pFLAG/HA-Ago2 by PCR using primers (5 -TGAGAA TTCAGTGGATCCACTAGTAACGG-3 and 5 -GCGG HeLa cells were cultured in Dulbecco’s Modified Ea- CCGCTAGCGTAATCGGGCACG-3 for pFLAG/HA, gle Medium (DMEM; Invitrogen) with 10% fetal bovine ◦ 5 -CATGGCGGCGGCGATATCGATCCG-3 and 5 - serum (FBS; Sigma) at 37 C with 5% CO2. Transfec- TACTCGGGAGCCGGCCCCGCACTTG-3 for pAgo2, tion was conducted using Lipofectamine 2000 (Invitrogen) and 5 -GCCGAGGTCAAGCGCGTGGGAGAC-3 and or polyethylenimine MAX (Polysciences) (29). For Lepto- 5 - TTCCACGGGCGTCTTGCCGGGCAGGATG-3 for mycin B (LMB; Calbiochem) treatment, 2 days after trans- pFLAG/HA-Ago2-Y529E), and self-ligated. fection, HeLa cells were washed twice with DMEM with FLAG/HA-Dcp1 expression plasmid was constructed 10% FBS and incubated in DMEM with 10% FBS contain- as follows: At first, both strands of chemically synthesized ing 10 ng/ml LMB or the same volume of solvent for 4 or 8 oligonucleotides for FLAG and HA tags (5 - CTAGCC h. CACCATGGACTACAAGGACGACGATGACAAGT ACCCTTATGACGTGCCCGATTACGCTA-3 and 5 - Fluorescence microscopy AGCTTAGCGTAATCGGGCACGTCATAAGGGTA CTTGTCATCGTCGTCCTTGTAGTCCATGGTGGG- HeLa cells were inoculated on a glass coverslip in a well 3 ) were annealed and inserted into NheI and HindIII of a 12-well culture plate at 1 × 105 cells/ml/well at 1 sites in pcDNA3.1-myc-Dicer (28) to generate pcDNA3.1- day before transfection. Total amount of transfected plas- FLAG/HA-Dicer. A fragment of Dcp1 CDS region mids per well was 0.5 or 1.0 g/well in any combinations was then amplified using cDNA from HEK293 cells by of used plasmids. At 2 days after transfection, cells were primers (5 -AAAGGGCTCGAGATGGAGGCGCTG washed with phosphate-buffered saline (PBS) and fixed AGTCGAGCTG-3 and 5 -AAAAGGAAAAGCGGCC with 4% paraformaldehyde at room temperature for 15 min. GCTCATAGGTTGTGGTTGTCTTTGTTC-3 ), and a The antibody staining was essentially performed as previ- FLAG/HA-containing vector fragment was amplified ously reported (30). Briefly, the fixed cells permeabilized 9858 Nucleic Acids Research, 2015, Vol. 43, No. 20 with PBS containing 0.2% Triton X-100 were incubated Cell fractionation with the first antibody (1:400–800 dilution) in PBS, and Nuclear and cytoplasmic fractions were isolated using this was followed by treatment with the second antibody NE-PER Nuclear and Cytoplasmic Extraction Reagent (1:400 dilution) in PBS. The cells were mounted in Pro- (Thermo scientific) according to the manufacturer’s recom- Long Gold antifade reagent with DAPI (Invitrogen) and mendation. observed under a Zeiss Axiovert 200 fluorescence micro- scope, using a ×63 objective lens. Quantification of fluores- cence microscopy images was performed using ImageJ soft- Immunoprecipitation ware (NIH). The first antibodies used for staining the cells Cells were plated in 9-cm dish at 2 × 106 cells/10 ml one day were as follows: anti-human AGO2 monoclonal antibody, before transfection. For the experiment of Supplementary 4G8 (Wako), Living Colors Full-Length Aequorea victoria Figure S6, 5 g of pFLAG/HA-FL or pFLAG/HA-Ago2 polyclonal GFP antibody (Clontech), mouse monoclonal and 5 g of pmyc-GFP-TNRC6A-NES-mut were trans- antibody to Dcp1a (Abcam; ab57654), anti-Rck/p54 poly- fected into the cells. Two days after transfection, cells were clonal antibody (MBL), anti-HA-tag antibody, 561 (MBL) washed once with PBS and lysed with ice-cold lysis buffer. and anti-HA-tag antibody, TANA2 (MBL). The second Lysates were centrifuged at 20 000 g for 10 min at 4◦C, and antibodies were as follows: FITC-conjugated goat anti- the supernatants were centrifuged at 48 000 g for 30 min rabbit IgG (Jackson Immunoresearch Laboratories), FITC- at 4◦C. Thirty l of protein A agarose (MILLIPORE) was conjugated goat anti-mouse IgG (Jackson Immunoresearch equilibrated with 500 l of lysis buffer and incubated with Laboratories), Cy5-conjugated goat anti-rabbit IgG (Jack- 2.5 g of an anti-GFP antibody in lysis buffer for 1 h at 4◦C. son Immunoresearch Laboratories), and Mouse IgG, Cy5- Then the cell lysates were added to the antibody-bound pro- linked (from goat) (GE Healthcare). tein A agarose and incubated with constant rotation for 2 hat4◦C. The beads were washed twice with the lysis buffer containing 300 mM NaCl, once with lysis buffer, and resus- pended in 40 lof2× SDS sample buffer. Bound proteins were eluted by boiling the beads for 5 min and separated by Western blot SDS-PAGE. For analyses of Ago-bound miRNAs, 10 gof Cells were lysed with ice-cold lysis buffer (10 mM Hepes– pFLAG/HA, pFLAG/HA-Ago2-WT or -Y529E were NaOH [pH 7.9], 1.5 mM MgCl2,10mMKCl,140mM transfected into the cells. Two days after transfection, cells NaCl, 0.5 mM DTT, 1 mM EDTA, 1 mM Na3VO4,10 lysates were prepared as described above and incubated mM NaF, 0.5% NP40, and 1× complete protease inhibitor with 60 l ANTI-FLAG M2 affinity gel (Sigma) with cocktail [Roche]) on ice for 30 min and centrifuged. Pro- constant rotation for 2 h at 4◦C. The beads were washed tein concentration of the supernatant was measured using twice with the lysis buffer containing 300 mM NaCl and a Bio-Rad protein assay kit with bovine serum albumin once with the lysis buffer. Sixty percent of beads were as the standard. Samples were separated by SDS-PAGE used for RNA extraction, and the remaining beads were and the gels were transferred to PVDF membranes us- used for protein analysis. Bound proteins were eluted by ing a Trans-Blot Turbo Transfer System (Bio-Rad). The boiling the beads for 5 min and separated by SDS-PAGE. membranes were blocked for 1 h in TBST (20 mM Tris– Quantification of Ago-bound miRNAs was performed as HCl [pH 7.5], 150 mM NaCl, and 0.1% Tween-20) sup- previously reported (10). plemented with 5% skim milk (Difco) and washed once For the experiment of Figure 6G, 2 g of pFLAG/HA, with TBST. Following incubation with the first antibody / ◦ pFLAG HA-Ago2-WT or -Y529E along with 8 gof (1:500–2000 dilution) in TBST at 4 C overnight, the mem- the expression construct of myc-GFP tagged proteins were branes were washed three times with TBST and treated transfected into the cells. Two days after transfection, cells with the second antibody (1:10 000 dilution) in TBST. Af- lysates were prepared as described above and incubated ter being washed three times with TBST, the membranes with 60 l ANTI-FLAG M2 affinity gel with constant rota- were incubated with ECL Prime Western Blotting Detec- tion for 2 h at 4◦C. The beads were washed three times with tion Reagent (GE Healthcare) and signals were detected by the lysis buffer, and resuspended in 40 lof2× SDS sample an ImageQuant LAS 4000 (GE Healthcare). The first anti- buffer. bodies used for western blot were as follows: anti-human AGO2 monoclonal antibody, 4G8, Living Colors Full- Transfection of small RNA duplexes Length Aequorea victoria polyclonal GFP antibody, mono- clonal ANTI-FLAG M2 antibody (Sigma), DYKDDDDK Fifty picomol/well of small RNA duplexes were transfected Tag Antibody (Cell Signaling Technology), mouse anti- into HeLa cells simultaneously or one day after transfection lamin A/C monoclonal antibody (BD Transduction Labo- with 0.5 g/well of pmyc-GFP-TNRC6A-NES-mut. RNA ratories), rabbit anti-GW182 (TNRC6A) antibody (Bethyl; oligonucleotides purchased from Sigma were as follows: A302-329A), and monoclonal anti-alpha tubulin antibody siControl (5 -CGUACGCGGAAUACUUCGAUU-3 and (ICN/CAPPEL Biomedicals). The second antibodies were 5 -UCGAAGUAUUCCGCGUACGUG-3 ); miR-200b as follows: ECL anti-mouse IgG, horseradish peroxidase (5 -UAAUACUGCCUGGUAAUGAUGA-3 and 5 - linked whole antibody (GE Healthcare) and ECL anti- CAUCUUACUGGGCAGCAUUGGA-3 ); and let-7b rabbit IgG, horseradish peroxidase linked whole antibody (5 -UGAGGUAGUAGGUUGUGUGGUU-3 and 5 - (GE Healthcare). CUAUACAACCUACUGCCUUCCC-3 ). Nucleic Acids Research, 2015, Vol. 43, No. 20 9859
Gene silencing activity assay sion condition of FLAG/HA-Ago1, -Ago2, or -Ago3 (Fig- ure 1B–E). In >80% of the cells expressing FLAG/HA- HeLa cells were plated at 1.0 × 105 cells/well in a well Ago1, -Ago2, or -Ago3, GFP-signal of TNRC6A-NES-mut of a 24-well plate at 1 day before transfection. Cells were was observed exclusively in the cytoplasmic foci with Ago simultaneously transfected with psiCHECK-CXCR4-PM proteins (Figure 1B–E). Expression level of FLAG/HA- or -MM 4× (10 ng/well), pGL3-Control (100 ng/well; Ago3 was apparently lower than that of FLAG/HA-Ago1 Promega) encoding firefly luciferase gene, /100 ng well or -Ago2 (Figure 1F), but it might be enough for inhibiting of shRNA expression construct (pSUPER-CXCR4 or the nuclear transport of TNRC6A-NES-mut protein (Fig- pSUPER-DsRed) or 50 pmol/well of siRNA (siCXCR4 ure 1E). The expression level of FLAG/HA-Ago4 was too or control siGY441 or siDsRed), the indicated amount of low to evaluate the effect of Ago4 overexpression on the pFLAG/HA or pFLAG/HA-Ago2, pmyc-GFP and/or subcellular localization of TNRC6A (data not shown). Fur- pmyc-GFP-TNRC6A or its mutant expression constructs. thermore, GFP signal of TNRC6A-NES-mut protein was The cells were harvested 48 h after the transfection, and observed exclusively in the cytoplasm in most of cells un- the relative Renilla/firefly luciferase activity was deter- der the overexpression condition of Ago2 proteins without mined using a Dual-Luciferase Reporter Assay System FLAG/HA-tag (Supplementary Figure S1), suggesting that (Promega). RNA sequences of siRNAs were as follows: Ago2 protein itself has certain ability to tether TNRC6A- siCXCR4 (5 -UGUUAGCUGGAGUGAAAACUU- NES-mut in the cytoplasm even when deleting FLAG/HA- 3 and 5 -GUUUUCACUCCAGCUAACACA-3 ); tag. In addition, we also reported that the translocation siGY441 (5 -AUGAUAUAGACGUUGUGGCUG-3 of TNRC6A protein from nucleus to cytoplasm is inhib- and 5 -GCCACAACGUCUAUAUCAUGG-3 ); siD- ited by the treatment with LMB, an inhibitor of Exportin sRed (5 -UUCUUCUGCAUUACGGGGCGG -3 and 1-dependent nuclear export of TNRC6A (10). In accor- 5 -GCCCCGUAAUGCAGAAGAAGA-3 ). dance with such results, when HeLa cells were treated with LMB, clear nuclear accumulation of the dotted GFP-signal RESULTS of myc-GFP-TNRC6A-WT protein was observed (Supple- mentary Figure S2B and F). However, in consistent with Overexpression of Ago proteins inhibits the nuclear transport the result using myc-GFP-tagged TNRC6A-NES-mut (Fig- of TNRC6A ure 1C and E), the nuclear transport of TNRC6A-WT Previously we reported that human TNRC6A is a nuclear– was inhibited by Ago2 overexpression (Supplementary Fig- cytoplasmic shuttling protein, and its subcellular localiza- ure S2C and F). These results clearly indicated that Ago2 tion is intrinsically regulated by its own NLS and NES overexpression inhibits the nuclear transport of myc-GFP- (10). In accordance with this finding, TNRC6A protein with TNRC6A-WT or myc-GFP-TNRC6A-NES-mut. NES mutation (TNRC6A-NES-mut) was predominantly We further examined the detailed subcellular distribution localized in the nucleus in human HeLa cells (10). How- of the ‘diffused’ TNRC6A proteins. For this purpose, we ever, when Ago proteins were overexpressed with TNRC6A- used anti-GFP antibody, which can detect the weak dif- NES-mut, we noticed that majority of TNRC6A-NES-mut fused GFP signals of myc-GFP-TNRC6A-NES-mut with proteins were colocalized with Ago proteins in the cyto- the higher sensitivity (Figure 2AandB;10), and classi- plasm, and rarely localized in the nucleus (Figure 1). The fied the cells into three groups according to the subcellu- result suggests that Ago proteins substantially govern the lar localization of the diffused signals (Figure 2C). In al- subcellular localization of TNRC6A. most all cells expressing control FLAG/HA-FL, most of To investigate the effect of Ago overexpression on the the diffused signals of myc-GFP-TNRC6A-NES-mut pro- subcellular localization of TNRC6A-NES-mut protein, 0.5 teins were observed in the nucleus in addition to the dot- g/well of the expression construct of myc-GFP-tagged tedfoci(Figure2A and C). In Ago2 overexpression condi- TNRC6A-NES-mut (pmyc-GFP-TNRC6A-NES-mut) was tion, most of the dotted foci of myc-GFP-TNRC6A-NES- transfected into HeLa cells with 0.5 g/well of the ex- mut proteins were also detected in the cytoplasm, and those pression construct of FLAG- and HA-tagged four types in the nucleus were rarely observed (Figure 2B and C). In of Ago proteins (pFLAG/HA-Ago1–Ago4) or control fire- about 85% of the cells, the diffused signals of TNRC6A- fly luciferase (pFLAG/HA-FL) (Figure 1). GFP signals of NES-mut proteins in the cytoplasm were stronger than or myc-GFP-TNRC6A-NES-mut in the cells were observed equal to those in the nucleus (Figure 2B and C). Consis- under a fluorescent microscope (Figure 1A–D), and cells tent with these results, the nuclear transport of the diffused were classified into three groups according to the localiza- myc-GFP-TNRC6A proteins induced by the LMB treat- tions of the ‘dotted foci’ with GFP signals at first, that ment was also evidently suppressed in Ago2 overexpression is, those with the dotted GFP signals (i) exclusively in the condition (Supplementary Figure S3). Thus, these results nucleus, (ii) exclusively in the cytoplasm, and (iii) in both indicated that the nuclear transport of the diffused myc- the nucleus and the cytoplasm (Figure 1E). When control GFP-TNRC6A proteins is inhibited by the overexpression FLAG/HA-FL was coexpressed with myc-GFP-TNRC6A- of Ago2 protein as well as that of the dotted myc-GFP- NES-mut, the dotted GFP-signal of TNRC6A-NES-mut TNRC6A proteins. protein was observed exclusively in the nucleus in 80–90% Next we examined whether overexpression of Ago2 pro- of the cells (Figure 1A and E), as observed in the cells tein inhibits the nuclear transport of endogenous TNRC6A. exogenously expressing TNRC6A-NES-mut protein alone However, since the amount of endogenous TNRC6A pro- (10). In contrast, such a nuclear-specific localization of teins might be too small to detect by immunohistochemical TNRC6A-NES-mut was not observed under overexpres- procedures using anti-TNRC6A antibody, we carried out 9860 Nucleic Acids Research, 2015, Vol. 43, No. 20
Figure 1. Overexpression of Ago proteins inhibits the nuclear transport of the dotted foci of GFP-signals of myc-GFP-TNRC6A-NES-mut proteins. (A–D) HeLa cells expressing myc-GFP-TNRC6A-NES-mut with either of the indicated proteins, FLAG/HA-FL (A), -Ago1 (B), -Ago2 (C), or -Ago3 (D), were stained with an anti-HA antibody, followed by Cy5-conjugated anti-mouse IgG. Fluorescent images are shown from the left: GFP signals of myc-GFP- TNRC6A-NES-mut; Cy5 signals of the second antibody against anti-HA antibody; the merged images of GFP and Cy5, in which GFP is shown in green, the Cy5 in magenta; the merged images with DAPI (blue). Bars, 20 m. (E) The ratio of cells expressing the dotted GFP-signals of myc-GFP-TNRC6A- NES-mut exclusively in the nucleus (N, blue), cytoplasm (C, red), or both (N+C, yellow). (F) Western blot of myc-GFP-TNRC6A-NES-mut and control FLAG/HA-FL, or FLAG/HA-tagged Ago1, Ago2, and Ago3 proteins. An anti-tubulin antibody was used as a loading control. Nucleic Acids Research, 2015, Vol. 43, No. 20 9861
Figure 2. Overexpression of Ago proteins inhibits the nuclear transport of the diffused signals of myc-GFP-TNRC6A-NES-mut and endogenous TNRC6A proteins. (A and B) HeLa cells expressing myc-GFP-TNRC6A-NES-mut with FLAG/HA-FL (A) or -Ago2 (B) were stained with an anti-GFP antibody and an anti-HA antibody, followed by FITC-conjugated anti-rabbit IgG and Cy5-conjugated anti-mouse IgG. Fluorescent images are shown from the left: FITC signals of the second antibody against anti-GFP antibody; Cy5 signals of the second antibody against anti-HA antibody; the merged images of FITC and Cy5, in which FITC is shown in green, the Cy5 in magenta; the merged images with DAPI (blue). Bars, 20 m. (C) The ratio of cells in which the diffused GFP signals of myc-GFP-TNRC6A-NES-mut in the nucleus were stronger than (N>C, blue), equal to (N = C,yellow),orweakerthan (N
Figure 3. Effects of different expression levels of Ago2 on the subcellular localization of TNRC6A-NES-mut protein. (A) Upper four panels, HeLa cells were transfected with pmyc-GFP-TNRC6A-NES-mut (0.5 g) and the mixture of pFLAG/HA-Ago2 and pFLAG/HA-FL (total amount = 0.5 g). Cell lysates were analyzed by western blot. Note that endogenous Ago2 (∼97 kDa) and FLAG/HA-Ago2 (∼100 kDa) proteins were detected as one overlapped band in the first panel, since they could not be separated in a 4–20% gel. Lowest panel, the ratio of cells expressing GFP-signal of myc-GFP-TNRC6A-NES- mut exclusively in the nucleus (N, blue), cytoplasm (C, red), or both (N+C, yellow). Typical fluorescent microscopy images are shown in Supplementary Figure S4. (B) Estimation of Ago2 protein level by fluorescent microscopy images. HeLa cells were transfected with pmyc-GFP-TNRC6A-NES-mut (0.5 g) and the mixture of pFLAG/HA-Ago2 and pFLAG/HA-FL (total amount = 0.5 g). Cells were stained with an anti-Ago2 antibody. Cells, in which no GFP signals were observed, were categorized as undetected (U), and others were classified by the subcellular localization of GFP-signal of myc-GFP- TNRC6A-NES-mut; exclusively in the nucleus (N), cytoplasm (C), or both (N+C). Data represents the mean ratio of Ago2 signal intensities ± standard deviation. The average signal intensity of Ago2 in the cells in which no GFP signals were not detected in the sample without transfection of pFLAG/HA- Ago2 (pFLAG/HA-Ago2, 0 mg, U) was set to 1. *P < 0.0001, t-test. n, the number of quantified cells. Typical fluorescent microscopy images are shown in Supplementary Figure S5. 9864 Nucleic Acids Research, 2015, Vol. 43, No. 20
Figure 4. TNRC6A is anchored in the cytoplasmic foci via its Ago-binding GW motifs. (A) Schematic representation of the domain structure of TNRC6A and its mutant proteins. GW-I, -II, and -III (black boxes), GW-repeated Ago-binding motifs. Q-rich (a green box), a glutamine-rich region. PAM2 (a magenta box), a poly(A)-binding protein binding motif 2 (PAM2). RRM (an orange box), RNA recognition motif. (B) The ratio of cells expressing GFP-signal of the indicated myc-GFP-TNRC6A mutant proteins exclusively in the nucleus (N, blue), cytoplasm (C, red), or both (N+C, yellow). (C–E) Fluorescent microscopy images of HeLa cells expressing the indicated myc-GFP-TNRC6A mutant protein with FLAG/HA-Ago2. Cells were stained with an anti-HA antibody. Bars, 20 m. Nucleic Acids Research, 2015, Vol. 43, No. 20 9865
GW-I and -II interact with Ago2 with strong affinities com- GFP-TNRC6A is colocalized with P body marker pro- paredtoGW-III(9,10). A TNRC6A mutant lacking both teins, Dcp1 and RCK/p54 (10). To investigate whether the GW-I and -II motifs (TNRC6A-del.GW-I+II) dramatically foci containing both TNRC6A-NES-mut and FLAG/HA- reduces the interaction with Ago2, and its mutant lacking Ago proteins are P bodies, the cells transfected with the three GW motifs (TNRC6A-del.GW-I+II+III) shows no or expression plasmids of myc-GFP-TNRC6A-NES-mut and little binding to Ago2 (10). FLAG/HA-Ago2 were stained by the antibody of Dcp1 At first, to investigate whether interaction of TNRC6A or RCK/p54 (Supplementary Figure S7). All of the foci with Ago2 in the normal condition without Ago2 over- containing myc-GFP-TNRC6A-NES-mut proteins were re- expression is saturated or not, the amounts of Ago2 pro- vealed to be colocalized with FLAG/HA-Ago2, Dcp1 or tein associated with TNRC6A-NES-mut in the normal and RCK/p54, suggesting that the cytoplasmic foci containing Ago2 overexpression conditions in HeLa cells were de- both TNRC6A-NES-mut and Ago2 are P bodies. termined by immunoprecipitation. Under the Ago2 over- Next, we tested the effect of Dcp1 or RCK/p54 overex- expression condition, more abundant amount of Ago2 pression on the subcellular localization of TNRC6A-NES- protein was immunoprecipitated with TNRC6A-NES-mut mut. In any cells transfected with Dcp1 or RCK/p54 ex- compared to the normal condition without Ago2 overex- pression construct, the exclusive cytoplasmic localization of pression (Supplementary Figure S6), suggesting that the GFP-signal of TNRC6A-NES-mut was not observed (Fig- amount of TNRC6A-interacting Ago2 protein is not sat- ure 5). In this condition, endogenous Ago2 was localized urated and more abundant Ago2 protein can interact with in the cytoplasmic foci with Dcp1 and RCK/p54 (date not TNRC6A. shown). The result suggests that Dcp1 or RCK/p54 is not Then, to investigate that the interaction between able to inhibit the nuclear transport of TNRC6A unlike Ago TNRC6A and Ago via GW motifs is required for the an- proteins. choring TNRC6A in the cytoplasm by Ago overexpression, we observed the subcellular localization of myc-GFP- Small RNA binding-deficient Ago2 is not able to anchor TNRC6A-NES-mut lacking each or combinations of TNRC6A protein in the cytoplasm GW-I, -II, and -III motifs under overexpression condition of Ago2. In the cells coexpressing myc-GFP-TNRC6A- The amount of Ago2 protein interacted with TNRC6A was NES-mut proteins deleting GW-I, -II, -III, -II+III, or revealed to be increased by Ago2 overexpression (Supple- -I+III (myc-GFP-TNRC6A-del.GW-I, -II, -III, -II+III, or mentary Figure S6), and TNRC6A is considered to be teth- -I+III-NES-mut) and FLAG/HA-Ago2, GFP signals were ered via interaction with Ago2 in the cytoplasmic P bodies mainly observed in the cytoplasm in more than 90% of the (Figure 4 and Supplementary Figure S2). Thus, it was spec- cells (Figure 4A-C). However, when pmyc-GFP-TNRC6A- ulated that Ago2 proteins are anchored to the P bodies by del.GW-I+II-NES-mut, the expression construct of unidentified interacting molecule(s) other than TNRC6A, myc-GFP-TNRC6A-NES-mut protein deleting both GW-I which could assemble Ago2 proteins as foci. One of the pos- and -II, was transfected with pFLAG/HA-Ago2, GFP sible mechanisms may be small RNA-mediated interaction signals were detected in the cytoplasm in ∼50% of the cells with other molecules, such as mRNAs. To investigate such (Figure 4B and D). Furthermore, myc-GFP-TNRC6A- possibility, we used Ago2 mutant (Ago2-Y529E), which is del.GW-I+II+III-NES-mut (Figure 4B and E), in which shown to be impaired in small RNA binding and localiza- all of three GW motifs, GW-I, -II, and -III, were deleted, tion in P body by the substitution of tyrosine to glutamic were extremely observed in the nucleus in about 90% of acid at position 529 amino acid, which is located in the bind- the cells, even when FLAG/HA-Ago2 was coexpressed. ing pocket of Ago2 with 5 -end of small RNA (31,32). We Similar results were observed by the transfection of pmyc- purified RNAs from immunoprecipitates of the cells trans- GFP-TNRC6A lacking each or combinations of GW-I, fected with wild-type Ago2 or Ago2-Y529E expression -II, and -III motifs with pFLAG/HA-Ago2, followed by constructs (pFLAG/HA-Ago2-WT or pFLAG/HA-Ago2- LMB treatment (Supplementary Figure S2A and D–F). Y529E), and performed RT-PCR for measuring the amount These results indicate that TNRC6A can be anchored in of typical miRNAs, miR-21 and let-7g. As previously de- the cytoplasm via GW motif(s) under the overexpression scribed, the amount of miRNAs associated with Ago2- condition of Ago2. And at least one of three GW motifs Y529E was low at less than 10% compared to that with is enough for tethering TNRC6A protein in Ago2 overex- Ago2-WT (Figure 6A and B), and Ago2-Y529E was not pression condition, although the ability of GW-III to tether localized in the P bodies (Supplementary Figure S8). Un- TNRC6A in the cytoplasm is weaker than those of others der overexpression condition of Ago2-Y529E, GFP-signal probably due to its weak binding ability to Ago2 protein. of TNRC6A-NES-mut was observed exclusively in the nu- cleus in >80% cells, although TNRC6A-NES-mut was pre- dominantly localized in the cytoplasm when the expression Cytoplasmic TNRC6A and Ago2 protein are co-localized level of Ago2-WT was similar to that of Ago2-Y529E (Fig- with P body markers ure 6C–F). Although immunoprecipitation experiments re- In the cells expressing abundant amount of Ago pro- vealed that the interaction of TNRC6A-NES-mut with teins, TNRC6A-NES-mut was co-localized with Ago pro- Ago2-Y529E was evidently reduced compared to that with teins in the cytoplasmic foci (Figure 1B–E). The size of Ago2-WT, this interaction was slightly stronger than the such foci increased according to the amount of transfected interaction between TNRC6A-del.GW-I+II-NES-mut and Ago2 expression constructs (Supplementary Figures S1A– Ago2-WT (Figure 6G). Furthermore, in the absence of C and S4). Previously we showed that cytoplasmic myc- TNRC6A overexpression, the overexpression of Ago2-WT 9866 Nucleic Acids Research, 2015, Vol. 43, No. 20
Figure 5. Overexpression of Dcp1 or RCK/p54 does not inhibit the nuclear transport of TNRC6A. (A and B) Fluorescent microscopy images of HeLa cells expressing myc-GFP-TNRC6A-NES-mut along with FLAG/HA-Dcp1 (A) or FLAG/HA-RCK/p54 (B). Cells were stained with an anti-HA antibody, followed by Cy5-conjugated anti-mouse IgG. Bars, 20 m. (C) The ratio of cells expressing GFP-signal of myc-GFP-TNRC6A-NES-mut exclusively in the nucleus (N, blue), cytoplasm (C, red), or both (N+C, yellow). (D) Western blot of myc-GFP-TNRC6A-NES-mut and FLAG/HA-tagged proteins. An anti-tubulin antibody was used as a loading control. formed large P-bodies but Ago2-Y529E could not form toplasmic anchoring of TNRC6A by Ago2 overexpression such P-bodies (Supplementary Figure S8). Thus, it is con- may require small RNAs loaded on Ago2 proteins. sidered that predominant nuclear localization of TNRC6A- NES-mut under overexpression condition of Ago2-Y529E Transfection of small RNAs does not affect the subcellular should be mainly due to impaired binding of Ago2-Y529E localization of TNRC6A-NES-mut to small RNAs, although a negligible effect of weak inter- action between TNRC6A-NES-mut and Ago2-Y529E may Above result indicates that interactions between Ago2 and be included. Thus, our results strongly suggest that the cy- small RNAs are important for the anchoring of TNRC6A in the cytoplasmic P bodies. Therefore, we investigated whether transfection of small RNAs affect the subcellular Nucleic Acids Research, 2015, Vol. 43, No. 20 9867
Figure 6. Overexpression of small RNA binding-deficient Ago2 protein does not Inhibit the nuclear transport of ATNRC6A.( ) Cell lysates from HeLa cells transfected with empty vector (pFLAG/HA), pFLAG/HA-Ago2-WT, or pFLAG/HA-Ago2-Y529E were immunoprecipitated with an anti-FLAG antibody. The immunoprecipitates (IP) and cell lysates (input) were analyzed by western blot. (B) MiRNAs contained in the immunoprecipitates of (A) was quantified by real-time PCR. Data are the mean and standard deviation of the triplicate measurements. (C and D) Fluorescent microscopy images of HeLa cells expressing myc-GFP-TNRC6A-NES-mut along with FLAG/HA-Ago2-WT (C) or -Y529E (D). Cells were stained with an anti-HA antibody. Bars, 20 m. (E) The ratio of cells expressing GFP-signal of myc-GFP-TNRC6A-NES-mut exclusively in the nucleus (N, blue), cytoplasm (C, red), or both (N+C, yellow). (F) Western blot of myc-GFP-TNRC6A-NES-mut and FLAG/HA-Ago2-WT or -Y529E. (G) Cell lysates from HeLa cells expressing myc-GFP- TNRC6A-NES-mut or myc-GFP-TNRC6A-NES-mut-del.GW-I+II along with FLAG/HA, FLAG/HA-Ago2-WT or -Y529E were immunoprecipitated with an anti-FLAG antibody. IPs and inputs were analyzed by western blot. 9868 Nucleic Acids Research, 2015, Vol. 43, No. 20 localization of TNRC6A-NES-mut. One day after transfec- manner (Figure 7E). The similar results were also obtained tion with pmyc-GFP-TNRC6A-NES-mut, small RNA du- by siRNA against CXCR4 (Supplementary Figure S11B). plexes were transfected into the cells (Supplementary Figure In addition, Ago2 overexpression did not affect the expres- S9). Transfection of siRNA or miRNA duplexes did not af- sion level of endogenous TNRC6A protein (Supplemen- fect the subcellular localization of GFP-signal of TNRC6A- tary Figure S10). This contradictory result for RNAi and NES-mut. Same results were obtained by the simultane- miRNA silencing by the overexpression of Ago2 proteins ous transfection with pmyc-GFP-TNRC6A-NES-mut and might be interpreted by the increased level of small RNA- small RNA duplexes (data not shown). These results sug- Ago2 complexes. When Ago2 protein is overexpressed, free gest that, since the intracellular machinery that processes small RNAs may be loaded on the newly synthesized Ago2 small RNAs is considered to be saturated (33,34), the trans- proteins. Such small RNA–Ago2 complexes in the cyto- fection of small RNAs does not increase the amount of plasm is known to act for enhancement of RNAi activity small RNAs loaded on Ago proteins in the cells. (35). However, since GW182 family proteins are necessary for miRNA silencing in addition to Ago proteins, actual miRNA silencing activities decreased by Ago2 overexpres- The counterbalance of the expression levels of Ago2 and sion. Consistent with this interpretation, when TNRC6A- TNRC6A proteins may modulate miRNA silencing activity WT or TNRC6A-NES-mut protein was co-expressed with Next we examined the effect of Ago2, TNRC6A or Ago2 protein, the activities of miRNA silencing were en- TNRC6A-NES-mut overexpression on the activities of hanced compared to the overexpression condition of Ago2 siRNA-based RNAi for mRNAs with perfectly comple- alone (Figure 7F). In addition, the necessity of interaction mentary sequences and miRNA silencing for mRNAs with between TNRC6A and Ago2 proteins for enhancement of partial complementarities. RNAi and miRNA silencing miRNA silencing activity was also confirmed by the result are considered to be performed through different silenc- that the transfection of expression constructs for TNRC6A- ing mechanisms: the former requires only Ago2 protein and del.GW-I+II+III or TNRC6A-del.GW-I+II+III-NES-mut small RNA in vitro (35) but the latter requires GW182 fam- could not increase the miRNA silencing activities (Fig- ily proteins in addition to Ago/small RNA (3,4). ure 7F). To analysis the respective activities, we used short hair- The overexpression of either of TNRC6A-WT or pin RNA (shRNA) against CXCR4 with Renilla luciferase TNRC6A-NES-mut, also attenuated miRNA silencing ac- reporter constructs containing a perfectly complementary tivity, although the attenuating effect was weaker for site or four bulged target sites, respectively, for CXCR4 TNRC6A-NES-mut than that for TNRC6A-WT (Fig- shRNA in its 3 UTR (Figure 7AandB;36). Consistent with ure 7E and F, and Supplementary Figure S11B), suggest- the previous reports (37,38), Ago2 overexpression enhanced ing that the event for reducing miRNA silencing activity RNAi activity in a dose dependent manner (Figure 7C), by TNRC6A overexpression may occur mainly in the cy- and such RNAi-enhancing effect by Ago2 was similarly or toplasm. TNRC6A is known to associate with various cy- slightly observed when TNRC6A-WT or any types of its toplasmic components essential for miRNA silencing, such mutant proteins were co-expressed (Figure 7D). Contrary, as PABPC1, PAN2-PAN3 and CCR4-NOT deadenylase overexpression of TNRC6A-WT weakened RNAi activity complexes. Most of the excessively synthesized TNRC6A dose-dependently (Figure 7C). Such RNAi-reducing effect proteins is also assumed to interact with these cytoplas- was evidently reduced by the overexpression of TNRC6A- mic components, but their interaction with Ago proteins NES-mut (Figure 7C), which is localized predominantly are over-saturated and all of the TNRC6A proteins in- in the nucleus. Overexpression of TNRC6A-WT or -NES- teracting with these components could not necessarily in- mut did not affect the expression level of endogenous teract with Ago proteins in HeLa cells. Such imbalance Ago2 protein (Supplementary Figure S10). These results of protein levels among Ago, TNRC6A, and other com- suggest that the reduction of RNAi activity may be de- ponents might result in the reduction of miRNA silenc- pendent on the cytoplasmic machinery. The similar re- ing activity. This was also speculated by the results of the sults were obtained by siRNA against CXCR4 (Supple- overexpression of TNRC6A proteins without Ago-binding mentary Figure S11A). When the empty vector of Ago2 motifs. The overexpression of TNRC6A-del.GW-I+II+III protein, pFLAG/HA, was transfected with TNRC6A-WT or TNRC6A-del.GW-I+II+III-NES-mut proteins also re- or TNRC6A-NES-mut expression constructs, unambigu- duced the miRNA silencing activities (Figure 7F), probably ous reduction of RNAi activity was also observed, al- because they also captured the TNRC6A-associated com- though the reduction level is rather low for TNRC6A- ponents. NES-mut compared to that for TNRC6A-WT (Figure 7D). However, such RNAi-reducing effects were marginal when DISCUSSION the TNRC6A proteins deleting Ago-binding motifs were overexpressed (Figure 7D; TNRC6A-del.GW-I+II+III and In this paper, we found that TNRC6A protein is tethered in TNRC6A-del.GW-I+II+III-NES-mut). These results may the cytoplasm under overexpression condition of Ago pro- indicate that the excessive amount of cytoplasmic TNRC6A teins (Figures 1–3 and Supplementary Figures S1–S5). Fur- proteins repress the RNAi activity by the endogenous Ago2 thermore, myc-GFP-tagged TNRC6A-WT or TNRC6A- proteins via direct or indirect interaction with Ago proteins NES-mut lacking all of three Ago binding GW-motifs was mainly in the cytoplasm. not able to be tethered in the cytoplasm and predomi- Unlike the effect on RNAi activity, Ago2 overexpression nantly localized in the nucleus even under overexpression attenuated miRNA silencing activity in a dose dependent condition of Ago2 (Figure 4 and Supplementary Figure Nucleic Acids Research, 2015, Vol. 43, No. 20 9869
Figure 7. TNRC6A overexpression modulates RNAi and miRNA silencing activities induced by shRNA. (A and B) The schematic base-pairing patterns between CXCR4 siRNA and Renilla luciferase mRNA harboring a perfectly complementary target site for CXCR4 siRNA (A) or four target sites with a central bulge (B) in its 3 UTR, which was encoded by psiCHECK-CXCR4-PM or -MM 4x, respectively. (C and E) HeLa cells were transfected with a mixture of the following four plasmids: psiCHECK-CXCR4-PM (C) or -MM 4× (E), pGL3-Control, pSUPER-CXCR4 or control pSUPER-DsRed, without or with 2.5, or 25, 250 ng/well of pFLAG/HA, pFLAG/HA-Ago2, pmyc-GFP-TNRC6A-WT, or pmyc-GFP-TNRC6A-NES-mut. Two day after transfection, luciferase activities were measured. The relative luciferase activity of pSUPER-CXCR4-transfected cells against that of control pSUPER- DsRed-transfected cells was set to ‘1’. The data were shown as the mean and standard deviation of the triplicate measurements. The actual values of relative luciferase activities of the control cells transfected with shRNA expression constructs alone were 3.4 ± 0.36% (C), and 2.1 ± 0.14% (E), respec- tively. (D and F) HeLa cells were transfected with a mixture of the following five plasmids: psiCHECK-CXCR4-PM (D) or -MM 4× (F), pGL3-Control, pSUPER-CXCR4 or -DsRed, 250 ng/well of pFLAG/HA or pFLAG/HA-Ago2, without or with 250 ng/well of the indicated TNRC6A expression con- structs. Two day after transfection, luciferase activities were measured. The relative luciferase activity of pSUPER-CXCR4-transfected cells against that of pSUPER-DsRed-transfected cells with pFLAG/HA transfection was set to ‘1’. The actual values of relative luciferase activities of the cells transfected with pFLAG/HA alone were 3.4 ± 0.29% (D), and 3.1 ± 0.23% (F), respectively.
S2), strongly suggested that the direct interaction between bition of nuclear transport of TNRC6A under overexpres- TNRC6A and Ago2 is essential for cytoplasmic anchor- sion condition of Ago, because TNRC6A mutants harbor- ing of TNRC6A by Ago2 overexpression. TNRC6A has ing only one GW-motif, either of GW-I or GW-II, were three independent Ago binding motifs (9,10), meaning that able to be tethered in the cytoplasm (Figure 4 and Supple- one molecule of TNRC6A protein is potentially able to mentary Figure S2). The result indicated that the binding bind three molecules of Ago proteins. However, multiple of one molecule of Ago protein with TNRC6A is neces- Ago-binding to TNRC6A was not necessary for the inhi- sary and enough for cytoplasmic localization of TNRC6A- 9870 Nucleic Acids Research, 2015, Vol. 43, No. 20
NES-mut protein. In our previous study (10), TNRC6A- NES-mut proteins were transported into the nucleus along with Ago2 proteins and miRNAs under normal condition in HeLa cells when Ago2 protein was not overexpressed (10). This might be contradictory paradox for this study. However, although we could not determine whether all TNRC6A-NES-mut proteins observed in the nucleus evi- dently associated with Ago2 proteins or not in our previous report (10), it was revealed that Ago2 proteins interacting with TNRC6A-NES-mut proteins increased by Ago2 over- expression in this study (Supplementary Figure S6). This means that free TNRC6A-NES-mut proteins without Ago2 interaction abundantly exist in the cells without Ago2 over- expression. Furthermore, Ago2-binding of TNRC6A was not essential for translocation of TNRC6A into the nucleus, because TNRC6A-del.GW-I+II+III-NES-mut can translo- Figure 8. Hypothetic model of contol of the subcellular localization of cate into the nucleus without interaction with Ago2 (10). TNRC6A proteins by Ago proteins. In human HeLa cells expressing a Thus, although a part of TNRC6A-NES-mut protein may small amount of Ago proteins (left), most of free TNRC6A proteins are be certainly translocated into the nucleus with Ago2 and distributed in the cytoplasm and shuttles between the nucleus and the cy- toplasm using its own NLS and NES, and a part of them may brings Ago- miRNAs, the remaining may be moved to the nucleus with- small RNA complexes into the nucleus from the cytoplasmic small P bod- out interaction with Ago2 proteins in the normal condi- ies. In the cells expressing a large amount of Ago proteins (right), they tion. While this manuscript was being revised, Schraivogel are strongly tethered in the P bodies probably through interaction between et al. reported that Ago proteins as well as TNRC6 pro- Ago-bound small RNAs and untranslating mRNAs, and enlarges the size of P bodies to form large foci. Most of TNRC6A proteins are then an- teins are able to shuttle between the cytoplasm and the nu- chored in the P bodies via its direct interaction with abundant amount of cleus and Ago expression level affects nuclear import of Ago proteins. TNRC6A (39). They demonstrated that TNRC6 proteins are imported into the nucleus by the Importin- pathway but Ago2 is not. Thus, their results advocate our previous Taken together, although Ago2 overexpression might not results that TNRC6A proteins are not necessarily translo- be the actual and ideal situation, the following model about cated into the nucleus with direct interaction with Ago2 pro- the anchoring of TNRC6A in the cytoplasmic P bodies by teins. Ago overexpression can be considered (Figure 8): excessive In the Ago2 overexpression conditions, the cytoplasmic amount of Ago proteins increases the size of P bodies by anchoring of TNRC6A proteins was revealed to be strongly their accumulation in the P bodies and is strongly tethered associated with Ago2 anchoring in the P bodies. TNRC6A- through interaction between Ago-bound small RNAs and NES-mut was colocalized with Ago proteins in the cyto- their target RNAs in the P bodies, and then TNRC6A is plasmic foci, which size increased according to the Ago anchored in the P bodies via its direct interaction with Ago. overexpression (Supplementary Figures S1 and S4). Such Schraivogel et al. showed that Ago overexpression inhibits cytoplasmic foci were marked by P body markers, such as the binding of Importin- to TNRC6A (39). This result Dcp1 and RCK/p54, even under Ago2 overexpression con- does not conflict with our model, because the previous re- dition (Supplementary Figure S7), indicating that such foci port demonstrated that Importin- is not colocalized with may be P bodies. However, overexpression of P body marker P bodies (41), suggesting that it might not be able to access proteins, Dcp1 and RCK/p54, could not anchor TNRC6A the P bodies where TNRC6A is anchored by Ago proteins. proteins in the cytoplasm and did not form P bodies sim- We showed that RNAi activity for cytoplasmic mRNAs ilar to those formed by Ago2 overexpression (Figure 5). by Ago2 overexpression enhanced RNAi activity in a dose Furthermore, the overexpression of TNRC6A did not form dependent manner (Figure 7C) in accordance with the pre- such large foci (10), demonstrating that the increasing of vious reports (37,38). For RNAi activity, since only Ago2 cytoplasmic P body size may mainly due to the accumula- protein and small RNA are necessary for induction of tion of increased amount of Ago proteins by their overex- RNAi (35), the absolute amount of Ago2 protein may reg- pression. It is known that untranslating mRNAs accumu- ulate RNAi activity in the cells. However, RNAi activity late in the P bodies (40), suggesting that the associations be- is inhibited by the overexpression of TNRC6A-WT more tween Ago2-bound small RNAs and the untranslating mR- severely than TNRC6A-NES-mut (Figure 7C and D). Since NAs in the P bodies might cause the Ago2 aggregation in TNRC6A-WT is predominantly localized in the cytoplasm P body. This was clearly indicated by the results of Ago2- but TNRC6A-NES-mut is in the nucleus (10), TNRC6A- Y529E protein (Figure 6). The overexpression of wild type WT may have a profound effect on the RNAi activity in Ago2 inhibited the nuclear transport of TNRC6A-NES- the cytoplasmic mRNA, in this case Renilla luciferase mR- mut protein, but that of Ago2-Y529E did not. The reason NAs. One possible mechanism for inhibition of RNAi in was considered to be the marginal amount of small RNAs the cytoplasm is inhibition of processing and/or Ago load- can loaded on Ago2-Y529E proteins, so that Ago2-Y529E ing of small RNA. It is reported that PIWI domains of could not sufficiently associate with untranslating mRNAs Ago proteins interacts with either GW182 family proteins in the P bodies (Figure 6). (5,8) or Dicer (42,43), which required for inducing RNAi by processing of shRNA into small RNA (44) and promoting Nucleic Acids Research, 2015, Vol. 43, No. 20 9871 small RNA loading onto Ago2 (45). The excessive amount 57). Therefore, it is considered that cytoplasmic tethering of of cytoplasmic TNRC6A may competitively disturb the in- TNRC6A by excessive amount of Ago proteins might occur teraction of Dicer with Ago2 and may result in the inhi- in some types of tissues or cell lines. bition of RNAi activity. However, siRNA but not shRNA One possible biological significance of cytoplasmic teth- against CXCR4, which is not necessary for Dicer cleavage, ering TNRC6A by Ago overexpression might promote the also showed the similar effects (Supplementary Figure S11), cytoplasmic function of Ago and/or TNRC6A or inhibit suggesting that the inhibition of RNAi effect is not due to nuclear function(s) of TNRC6A. In accordance with our the inhibition of direct interaction between Ago and Dicer results, recent reports have suggested that human GW182 by the overexpression of TNRC6A protein. Therefore, other proteins might function in the nucleus as well as in the cy- possible mechanism(s) of inhibition of RNAi and miRNA toplasm (10,22–24). However, precise function(s) of nuclear silencing induced by overexpression of TNRC6A protein is GW182 proteins are not well understood. Future analyses that the interaction between TNRC6A protein with Ago2 will uncover the significance of the regulation of TNRC6A protein inhibits RNAi and miRNA silencing by inhibit- subcellular localization. ing siRNA loading into Ago2 or interaction with target mRNA. SUPPLEMENTARY DATA In addition, we also showed that Ago2 overexpression at- tenuated miRNA silencing (Figure 7E and F). This result is Supplementary Data are available at NAR Online. consistent with the recent report which showed that Ago2 overexpression reduces off-targeting effects of siRNA (38), ACKNOWLEDGEMENT because the mechanism of siRNA-based off-target effect is We thank Drs Thomas Tuschl and Patrick Provost for considered to be similar to that of miRNA silencing (46– kindly providing plasmids and Ai Nishi for technical assis- 48). Not only Ago2 but also TNRC6A or TNRC6A-NES- tance of plasmid construction. mut overexpression attenuated miRNA silencing under nor- mal condition (Figure 7E and F). However, previous reports showed that TNRC6s overexpression enhanced miRNA si- FUNDING lencing (14,49). Such contradictory result may be explained Ministry of Education, Culture, Sports, Science and Tech- by our experiments that TNRC6A or TNRC6A-NES-mut nology of Japan (MEXT); Cell Innovation Project (MEXT overexpression enhanced miRNA silencing under overex- to K.U.-T.). Funding for open access charge: Ministry of pression condition of Ago2 whereas their overexpression Education, Culture, Sports, Science and Technology of reduced miRNA silencing activity under normal condition Japan (MEXT to K.U.-T.). (Figure 7E and F). Therefore, the relative expression lev- Conflict of interest statement. None declared. els between Ago and TNRC6 proteins might determine the effect of TNRC6 proteins on the miRNA silencing. These suggest that the counterbalance in the expression levels of REFERENCES Ago and GW182 family proteins might regulate miRNA 1. Ender,C. and Meister,G. (2010) Argonaute proteins at a glance. J. silencing activity: the excessive amount of either Ago or Cell Sci., 123, 1819–1823. 2. Meister,G. (2013) Argonaute proteins: functional insights and TNRC6 proteins may competitively inhibit function of Ago emerging roles. Nat. Rev. Genet., 14, 447–459. proteins and reduce the miRNA silencing activity.In our ex- 3. Eulalio,A., Tritschler,F. and Izaurralde,E. (2009) The GW182 protein perimental conditions, it is unambiguously revealed that the family in animal cells: New insights into domains required for proportions between Ago proteins and a human GW182 miRNA-mediated gene silencing. RNA, 15, 1433–1442. 4. Braun,J.E., Huntzinger,E. and Izaurralde,E. (2012) A molecular link protein, TNRC6A, in association with other components between miRISCs and deadenylases provides new insight into the may be important for regulating miRNA silencing activity. mechanism of gene silencing by microRNAs. Cold Spring Harb. The subcellular localization of one of the essential compo- Perspect. Biol., 4, a012328. nents for miRNA silencing, TNRC6A, may play a key part 5. Till,S., Lejeune,E., Thermann,R., Bortfeld,M., Hothorn,M., in modulation of silencing efficiently. However, as all our Enderle,D., Heinrich,C., Hentze,M.W. and Ladurner,A.G. (2007) A conserved motif in Argonaute-interacting proteins mediates results about RNAi and miRNA silencing were dependent functional interactions through the Argonaute PIWI domain. Nat. on the experiments of the overexpression of tagged proteins, Struct. Mol. Biol., 14, 897–903. we cannot exclude the possibility that these might not nec- 6. Eulalio,A., Helms,S., Fritzsch,C., Fauser,M. and Izaurralde,E. (2009) essarily be results from actual and ideal conditions. A C-terminal silencing domain in GW182 is essential for miRNA ∼ function. RNA, 15, 1067–1077. Our results showed that at least 2-fold increase of Ago2 7. Lazzaretti,D., Tournier,I. and Izaurralde,E. (2009) The C-terminal protein inhibited the nuclear transport of TNRC6A in domains of human TNRC6A, TNRC6B, and TNRC6C silence HeLa cells (Figure 3). It was reported that the amount of bound transcripts independently of Argonaute proteins. RNA, 15, Ago2 protein is about 60% of total cellular Ago family pro- 1059–1066. teins (Ago1∼4) in HeLa cells (50). Thus, ∼2-fold increase 8. Lian,S.L., Li,S., Abadal,G.X., Pauley,B.A., Fritzler,M.J. and ∼ Chan,E.K. (2009) The C-terminal half of human Ago2 binds to of Ago2 protein corresponds approximately 1.6-fold in- multiple GW-rich regions of GW182 and requires GW182 to mediate crease of total cellular Ago proteins, meaning that such silencing. RNA, 15, 804–813. small amount of increase might inhibit the nuclear trans- 9. Takimoto,K., Wakiyama,M. and Yokoyama,S. (2009) Mammalian port of TNRC6A if all Ago proteins had same activities GW182 contains multiple Argonaute-binding sites and functions in microRNA-mediated translational repression. RNA, 15, 1078–1089. for cytoplasmic anchoring of TNRC6A. The expression lev- 10. Nishi,K., Nishi,A., Nagasawa,T. and Ui-Tei,K. (2013) Human els of Agos differ in various types of tissues or cell lines, TNRC6A is an Argonaute-navigator protein for and are affected by tumorigenesis and cell signaling (27,51– microRNA-mediated gene silencing in the nucleus. RNA, 19, 17–35. 9872 Nucleic Acids Research, 2015, Vol. 43, No. 20
11. Fabian,M.R., Mathonnet,G., Sundermeier,T., Mathys,H., Rho family GTPases but dependent on dynamin and Arf6. J. Biol. Zipprich,J.T., Svitkin,Y.V., Rivas,F., Jinek,M., Wohlschlegel,J., Chem., 282, 27503–27517. Doudna,J.A. et al. (2009) Mammalian miRNA RISC recruits CAF1 31. R¨udel,S., Wang,Y., Lenobel,R., Korner,R.,¨ Hsiao,H.H., Urlaub,H., and PABP to affect PABP-dependent deadenylation. Mol. Cell, 35, Patel,D. and Meister,G. (2011) Phosphorylation of human Argonaute 868–880. proteins affects small RNA binding. Nucleic Acids Res., 39, 12. Fabian,M.R., Cieplak,M.K., Frank,F., Morita,M., Green,J., 2330–2343. Srikumar,T., Nagar,B., Yamamoto,T., Raught,B., Duchaine,T.F. 32. Pare,J.M., Lopez-Orozco,J.´ and Hobman,T.C. (2011) et al. (2011) miRNA-mediated deadenylation is orchestrated by MicroRNA-binding is required for recruitment of human Argonaute GW182 through two conserved motifs that interact with 2 to stress granules and P-bodies. Biochem. Biophys. Res. Commun., CCR4–NOT. Nat. Struct. Mol. Biol., 18, 1211–1217. 414, 259–264. 13. Zekri,L., Huntzinger,E., Heimstadt,S.¨ and Izaurralde,E. (2009) The 33. Khan,A.A., Betel,D., Miller,M.L., Sander,C., Leslie,C.S. and silencing domain of GW182 interacts with PABPC1 to promote Marks,D.S. (2009) Transfection of small RNAs globally perturbs gene translational repression and degradation of microRNA targets and is regulation by endogenous microRNAs. Nat. Biotechnol., 27, 549–555. required for target release. Mol. Cell. Biol., 29, 6220–6231. 34. Nagata,Y., Shimizu,E., Hibio,N. and Ui-Tei,K. (2013) Fluctuation of 14. Huntzinger,E., Braun,J.E., Heimstadt,S.,¨ Zekri,L. and Izaurralde,E. global gene expression by endogenous miRNA response to the (2010) Two PABPC1-binding sites in GW182 proteins promote introduction of an exogenous miRNA. Int. J. Mol. Sci., 14, miRNA-mediated gene silencing. EMBO J., 29, 4146–4160. 11171–11189. 15. Jinek,M., Fabian,M.R., Coyle,S.M., Sonenberg,N. and Doudna,J.A. 35. Rivas,F.V., Tolia,N.H., Song,J.J., Aragon,J.P., Liu,J., Hannon,G.J. (2010) Structural insights into the human GW182-PABC interaction and Joshua-Tor,L. (2005) Purified Argonaute2 and an siRNA form in microRNA-mediated deadenylation. Nat. Struct. Mol. Biol., 17, recombinant human RISC. Nat. Struct. Mol. Biol., 12, 340–349. 238–240. 36. Doench,J.G., Petersen,C.P. and Sharp,P.A. (2003) siRNAs can 16. Braun,J.E., Huntzinger,E., Fauser,M. and Izaurralde,E. (2011) function as miRNAs. Genes Dev., 17, 438–442. GW182 proteins recruit cytoplasmic deadenylase complexes to 37. Diederichs,S., Jung,S., Rothenberg,S.M., Smolen,G.A., Mlody,B.G. miRNA targets. Mol. Cell, 44, 120–133. and Haber,D.A. (2008) Coexpression of Argonaute-2 enhances RNA 17. Chekulaeva,M., Mathys,H., Zipprich,J.T., Attig,J., Colic,M., interference toward perfect match binding sites. Proc. Natl. Acad. Sci. Parker,R. and Filipowicz,W. (2011) miRNA repression involves U.S.A., 105, 9284–9289. GW182-mediated recruitment of CCR4-NOT through conserved 38. Borner,K.,¨ Niopek,D., Cotugno,G., Kaldenbach,M., Pankert,T., W-containing motifs. Nat. Struct. Mol. Biol., 18, 1218–1226. Willemsen,J., Zhang,X., Sch¨urmann,N., Mockenhaupt,S., Serva,A. 18. Eystathioy,T., Chan,E.K., Tenenbaum,S.A., Keene,J.D., Griffith,K. et al. (2013) Robust RNAi enhancement via human Argonaute-2 and Fritzler,M.J. (2002) A phosphorylated cytoplasmic autoantigen, overexpression from plasmids, viral vectors and cell lines. Nucleic GW182, associates with a unique population of human mRNAs Acids Res., 41, e199. within novel cytoplasmic speckles. Mol. Biol. Cell, 13, 1338–1351. 39. Schraivogel,D., Schindler,S.G., Danner,J., Kremmer,E., Pfaff,J., 19. Eystathioy,T., Jakymiw,A., Chan,E.K., Seraphin,B.,´ Cougot,N. and Hannus,S., Depping,R. and Meister,G. (2015) Importin- facilitates Fritzler,M.J. (2003) The GW182 protein colocalizes with mRNA nuclear import of human GW proteins and balances cytoplasmic degradation associated proteins hDcp1 and hLSm4 in cytoplasmic gene silencing protein levels. Nucleic Acids Res., 43, 7447–7461. GW bodies. RNA, 9, 1171–1173. 40. Parker,R. and Sheth,U. (2007) P bodies and the control of mRNA 20. Liu,J., Rivas,F.V., Wohlschlegel,J., Yates,JR III., Parker,R. and translation and degradation. Mol. Cell., 25, 635–646. Hannon,G.J. (2005) A role for the P-body component GW182 in 41. Chang,W.L. and Tarn,W.Y. (2009) A role for transportin in microRNA function. Nat. Cell Biol., 7, 1261–1266. deposition of TTP to cytoplasmic RNA granules and mRNA decay. 21. Franks,T.M. and Lykke-Andersen,J. (2008) The control of mRNA Nucleic Acids Res., 37, 6600–6612. decapping and P-body formation. Mol. Cell, 32, 605–615. 42. Doi,N., Zenno,S., Ueda,R., Ohki-Hamazaki,H., Ui-Tei,K. and 22. Ameyar-Zazoua,M., Rachez,C., Souidi,M., Robin,P., Fritsch,L., Saigo,K. (2003) Short-interfering-RNA-mediated gene silencing in Young,R., Morozova,N., Fenouil,R., Descostes,N., Andrau,J.C. et al. mammalian cells requires Dicer and eIF2C translation initiation (2012) Argonaute proteins couple chromatin silencing to alternative factors. Curr. Biol., 13, 41–46. splicing. Nat. Struct. Mol. Biol., 19, 998–1004. 43. Tahbaz,N., Kolb,F.A., Zhang,H., Jaronczyk,K., Filipowicz,W. and 23. Matsui,M., Chu,Y., Zhang,H., Gagnon,K.T., Shaikh,S., Hobman,T.C. (2004) Characterization of the interactions between Kuchimanchi,S., Manoharan,M., Corey,D.R. and Janowski,B.A. mammalian PAZ PIWI domain proteins and Dicer. EMBO Rep., 5, (2013) Promoter RNA links transcriptional regulation of 189–194. inflammatory pathway genes. Nucleic Acids Res., 41, 10086–10109. 44. Kanellopoulou,C., Muljo,S.A., Kung,A.L., Ganesan,S., Drapkin,R., 24. Gagnon,K.T., Li,L., Chu,Y., Janowski,B.A. and Corey,D.R. (2014) Jenuwein,T., Livingston,D.M. and Rajewsky,K. (2005) RNAi factors are present and active in human cell nuclei. Cell Rep., 6, Dicer-deficient mouse embryonic stem cells are defective in 211–221. differentiation and centromeric silencing. Genes Dev., 19, 489–501. 25. Kim,M.S., Oh,J.E., Kim,Y.R., Park,S.W., Kang,M.R., Kim,S.S., 45. Ota,H., Sakurai,M., Gupta,R., Valente,L., Wulff,B.E., Ariyoshi,K., Ahn,C.H., Yoo,N.J. and Lee,S.H. (2010) Somatic mutations and Iizasa,H., Davuluri,R.V. and Nishikura,K. (2013) ADAR1 forms a losses of expression of microRNA regulation-related genes AGO2 complex with Dicer to promote microRNA processing and and TNRC6A in gastric and colorectal cancers. J. Pathol., 221, RNA-induced gene silencing. Cell, 153, 575–589. 139–146. 46. Lin,X., Ruan,X., Anderson,M.G., McDowell,J.A., Kroeger,P.E., 26. Yoo,N.J., Hur,S.Y., Kim,M.S., Lee,J.Y. and Lee,S.H. (2010) Fesik,S.W. and Shen,Y. (2005) siRNA-mediated off-target gene Immunohistochemical analysis of RNA-induced silencing silencing triggered by a 7 nt complementation. Nucleic Acids Res., 33, complex-related proteins AGO2 and TNRC6A in prostate and 4527–4535. esophageal cancers. APMIS, 118, 271–276. 47. Jackson,A.L., Burchard,J., Schelter,J., Chau,B.N., Cleary,M., Lim,L. 27. Meister,G., Landthaler,M., Patkaniowska,A., Dorsett,Y., Teng,G. and Linsley,P.S. (2006) Widespread siRNA ‘off-target’ transcript and Tuschl,T. (2004) Human Argonaute2 mediates RNA cleavage silencing mediated by seed region sequence complementarity. RNA, targeted by miRNAs and siRNAs. Mol. Cell, 15, 185–197. 12, 1179–1187. 28. Provost,P., Dishart,D., Doucet,J., Frendewey,D., Samuelsson,B. and 48. Ui-Tei,K., Naito,Y., Nishi,K., Juni,A. and Saigo,K. (2008) Radmark,O.˚ (2002) Ribonuclease activity and RNA binding of Thermodynamic stability and Watson-Crick base pairing in the seed recombinant human Dicer. EMBO J., 21, 5864–5874. duplex are major determinants of the efficiency of the siRNA-based 29. Thomas,M., Lu,J.J., Ge,Q., Zhang,C., Chen,J. and Klibanov,A.M. off-target effect. Nucleic Acids Res., 36, 7100–7109. (2005) Full deacylation of polyethylenimine dramatically boosts its 49. Li,S., Wang,L., Fu,B., Berman,M.A., Diallo,A. and Dorf,M.E. gene delivery efficiency and specificity to mouse lung. Proc. Natl. (2014) TRIM65 regulates microRNA activity by ubiquitination of Acad. Sci. U.S.A., 102, 5679–5684. TNRC6. Proc. Natl. Acad. Sci. U.S.A., 111, 6970–6975. 30. Nishi,K. and Saigo,K. (2007) Cellular internalization of green 50. Petri,S., Dueck,A., Lehmann,G., Putz,N., R¨udel,S., Kremmer,E. and fluorescent protein fused with herpes simplex virus protein VP22 viaa Meister,G. (2011) Increased siRNA duplex stability correlates with lipid raft-mediated endocytic pathway independent of caveolae and reduced off-target and elevated on-target effects. RNA, 17, 737–749. Nucleic Acids Research, 2015, Vol. 43, No. 20 9873
51. Sasaki,T., Shiohama,A., Minoshima,S. and Shimizu,N. (2003) 55. Sand,M., Skrygan,M., Georgas,D., Arenz,C., Gambichler,T., Identification of eight members of the Argonaute family inthe Sand,D., Altmeyer,P. and Bechara,F.G. (2012) Expression levels of human genome small star, filled. Genomics, 82, 323–330. the microRNA maturing microprocessor complex component 52. Adams,B.D., Claffey,K.P. and White,B.A. (2009) Argonaute-2 DGCR8 and the RNA-induced silencing complex (RISC) expression is regulated by epidermal growth factor receptor and components argonaute-1, argonaute-2, PACT, TARBP1, and mitogen-activated protein kinase signaling and correlates with a TARBP2 in epithelial skin cancer. Mol. Carcinog., 51, 916–922. transformed phenotype in breast cancer cells. Endocrinology, 150, 56. Li,W., Liu,M., Feng,Y., Xu,Y.F., Che,J.P., Wang,G.C., Zheng,J.H. 14–23. and Gao,H.J. (2013) Evaluation of Argonaute protein as a predictive 53. Chang,S.S., Smith,I., Glazer,C., Hennessey,P. and Califano,J.A. marker for human clear cell renal cell carcinoma. Int. J. Clin. Exp. (2010) EIF2C is overexpressed and amplified in head and neck Pathol., 6, 1086–1094. squamous cell carcinoma. ORL J. Otorhinolaryngol. Relat. Spec., 72, 57. Yang,F.Q., Huang,J.H., Liu,M., Yang,F.P., Li,W., Wang,G.C., 337–343. Che,J.P. and Zheng,J.H. (2014) Argonaute 2 is up-regulated in tissues 54. Li,L., Yu,C., Gao,H. and Li,Y. (2010) Argonaute proteins: potential of urothelial carcinoma of bladder. Int. J. Clin. Exp. Pathol., 7, biomarkers for human colon cancer. BMC Cancer, 10, 38. 340–347.